How to pick the right transistor?

[[email protected]]

I have been trying to read up on formulas for selecting transistors. I
am a bit math impaired, but not too badly. I am totally confused about
what each tearm means. For instance the Ic and Hfe, I kind of
understand them, but not enough to do my own calculations for my
designs. I need an image with examples on picking the right base
resistor for the transistor. If I could rate my confusion from 1 to 10
, it would be around 8. Thanks to all in advance.

 

[Jim Land]

[email protected] wrote in @i3g2000cwc.googlegroups.com:

I have been trying to read up on formulas for selecting transistors. I
am a bit math impaired, but not too badly. I am totally confused about
what each tearm means. For instance the Ic and Hfe, I kind of
understand them, but not enough to do my own calculations for my
designs. I need an image with examples on picking the right base
resistor for the transistor. If I could rate my confusion from 1 to 10
, it would be around 8. Thanks to all in advance.

Picking transistors and values for components connected to them is an
advanced electronics skill. Some people go to college and get an
engineering degree to learn this.

Best for you to start by building circuits already designed by someone
else, using schematics you find in books or magazines or by Googling the
Web. (LOTS of them are available! You can find a circuit to do just
about anything you can think of.)

Once you've built some circuits and find they don't work right off, and
then got them working the way they're supposed to, you can start
understanding how these circuits work and why the components have the
values they do. You'll see why a certain transistor works well in a
partcular kind of circuit, and can try using it in a similar circuit.

And you can ask some intelligent questions about circuit design here.

 

[[email protected]]

[email protected] wrote in @i3g2000cwc.googlegroups.com:


Picking transistors and values for components connected to them is an
advanced electronics skill. Some people go to college and get an
engineering degree to learn this.

Best for you to start by building circuits already designed by someone
else, using schematics you find in books or magazines or by Googling the
Web. (LOTS of them are available! You can find a circuit to do just
about anything you can think of.)

Once you've built some circuits and find they don't work right off, and
then got them working the way they're supposed to, you can start
understanding how these circuits work and why the components have the
values they do. You'll see why a certain transistor works well in a
partcular kind of circuit, and can try using it in a similar circuit.

And you can ask some intelligent questions about circuit design here.

My particular situation is that I need to find a transistor that can
drive a 220mA load with a very high current gain, I was thinking maybe
a darlington pair, but there is just too many too choose from.

 

[default]

My particular situation is that I need to find a transistor that can
drive a 220mA load with a very high current gain, I was thinking maybe
a darlington pair, but there is just too many too choose from.

OK you have one paramater (current) and have some idea what gain you
need, presumably you also know what voltage it has to stand off - then
you'll just have to pick one of the hundreds that will work. As
likely as not, there won't be one transistor that will work, but many.
There probably won't be one that is clearly the best, unless your
application is tough.

What is the application?

Don't rule out mosfets. They are voltage operated rather than current
driven - hi input impedance translates to high gain since they take
pico amperes to turn on. (but many require about four volts of bias
to turn on if that is a concern) Very efficient in lots of
applications.

A mosfet that can switch 60 amps may have an "on" resistance of a few
hundredths of an ohm so it won't drop much voltage. The trade off
comes when you need to switch high voltage - then the "on" resistance
for a HV part is higher. At some point it makes more sense to use a
bipolar or an insulated gate bipolar (best of both worlds for some
applications)

 

[default]

I have been trying to read up on formulas for selecting transistors. I
am a bit math impaired, but not too badly. I am totally confused about
what each tearm means. For instance the Ic and Hfe, I kind of
understand them, but not enough to do my own calculations for my
designs. I need an image with examples on picking the right base
resistor for the transistor. If I could rate my confusion from 1 to 10
, it would be around 8. Thanks to all in advance.

These are a few of the popular transistors for high gain NPN in small
signal low voltage applications . . . Popular= common and low cost

Check out the 2N7000

It is a handy mosfet that can directly replace an NPN transistor in
lots of applications - with lots of gain. TO92 case

Takes about 4 volts to turn on with a few microamps of current -
pulsed operation with 2 amps of drain current and it takes 8 volts to
turn on

http://www.fairchildsemi.com/pf/2N/2N7000.html\ (propaganda)

http://www.fairchildsemi.com/ds/2N/2N7000.pdf (data sheet)

From the propaganda:
They can be used in most applications requiring up to 400mA DC and can
deliver pulsed currents up to 2A. These products are particularly
suited for low voltage, low current applications such as small servo
motor control, power MOSFET gate drivers, and other switching
applications.

and they are very inexpensive

The other old standby for high gain NPN is the MPSA13 darlington

www.onsemi.com/pub/Collateral/MPSA13-D.PDF

DC gain of 5,000+ collector current of 500 ma

inexpensive and very common TO92 case

 

[[email protected]]

Thanks default, I will look at the links you offerd. You mentioned
using MOSFETs, I may not be able to use these because I need to drive
two different base(gate) with less than 25mA from a PIC Output signal
that also goes through a voltage drop (due to a diode) to about 3.8v.
So the MOSFETS, without a transistor amplifier, are out of the
question.

 

[Michael Black]

I have been trying to read up on formulas for selecting transistors. I
am a bit math impaired, but not too badly. I am totally confused about
what each tearm means. For instance the Ic and Hfe, I kind of
understand them, but not enough to do my own calculations for my
designs. I need an image with examples on picking the right base
resistor for the transistor. If I could rate my confusion from 1 to 10
, it would be around 8. Thanks to all in advance.

In hobby circles, generally a small selection are used. People pick
what they are seeing in the other articles in the hobby magazines, and
those types are reinforced by what the hobby outlets carry. So somehow
decades ago, someone started using a 2N2222, and the hobby outlets decided
to carry it (it being a recognizable device), which fueled it's use in
further projects.

So traditionally if someone was starting from scratch rather than copying
an existing circuit, they'd go through the relatively short list of
devices that could be bought easily, checking the specs if they didn't
immediately notice one that stood out for their purpose, until they
found something suitable. Sure, there were loads of other devices
available, but maybe not in small quantities or without requiring lots
of gyrations.

I suppose that has changed with places like Mouser and Digikey, relatively
accessible to the hobbyist but a fairly full line catalog.

Michael

 

[John Popelish]

My particular situation is that I need to find a transistor that can
drive a 220mA load with a very high current gain, I was thinking maybe
a darlington pair, but there is just too many too choose from.

I would be looking for a transistor with an advertised maximum current
3 to 10 times higher than such a requirement, since most transistors
hit their peak current gain at 1/3 to 1/10th of maximum rated Ic
(collector current). But if you need highest gain, you might look at
transistors specifically designed with this characteristic, like the
ones made by Zetex. You don't mention whether you need PNP or NPN,
the highest collector voltage the device must withstand, whether you
are switching 220 mA or linearly amplifying at that current, the
signal frequency (or switching speed), how important low on state
voltage drop is (if switching). But if you are switching, remember
that darlingtons always drop at least 1.2 volts when on, whereas
simple transistors might saturate down to a small fraction of a volt
when on, lowering the on state losses.

Here are a pair of random Zetex data sheets to give you some idea what
you might find from them:

http://www.zetex.com/3.0/pdf/ZTX718.pdf
http://www.zetex.com/3.0/pdf/ZTX1049A.pdf

a curve you should think about (for switching) are VCE(sat) (collector
to emitter saturation voltage, with a specified ratio between
collector current and base current up to 50/1 or a base current of 4.4
mA for your load). Both these transistors keep the saturation voltage
well below .1 volt at 220 mA and 4.4 mA base current. If you can
tolerate a little more saturation voltage, you could use less base
current.

Also, look at the current gain curves versus collector voltage (Hfe
vs. Ic) to see that the peak gain is well below rated maximum
collector current.

Digikey carries these and many more, and their search engine can be
used to narrow down the choices by polarity, collector voltage and
current ratings and package, before you dive into data sheets.
http://www.digikey.com/

 

[Eeyore]

My particular situation is that I need to find a transistor that can
drive a 220mA load with a very high current gain, I was thinking maybe
a darlington pair, but there is just too many too choose from.

You choose a popular one !

Easily recognised from the stock levels held by distributors like Mouser etc

Graham

 

[Jonathan Kirwan]

Thanks default, I will look at the links you offerd. You mentioned
using MOSFETs, I may not be able to use these because I need to drive
two different base(gate) with less than 25mA from a PIC Output signal
that also goes through a voltage drop (due to a diode) to about 3.8v.
So the MOSFETS, without a transistor amplifier, are out of the
question.

At 25mA, the PIC outputs probably won't be 4.5V (assuming a 5V supply,
about which I'm also not sure.) My guess is that they will be about
25mA*70ohms or 1.75V less than your supply voltage before they meet
your diode. (Actual impedance will vary some and you need to be aware
that you cannot drive all pins at the same 25mA, as there is another
spec that provides an absolute maximum [don't get too close to it]
that limits the total port output.) Or were you thinking that it was
a diode there in the PIC to explain the 3.8V output?

Anyway, I think we understand the following:

Drive signal: PIC output
Number of circuits: 2
Mode of operation: switch (unless I assume far too much here)
Load current: 220mA
Load impendance: unknown, but may have significant L, if relay
Load voltage rail: unknown
PIC Voltage rail: ??, but either 4.5V (calculated from your
assumed diode drop and 3.8V mentioned) or
else 5V (blindly assumed as "normal")

Could you expand the above to provide more precise information? A
highly inductive or highly capacitive load may require other caveats
as compared to a purely resistive load. Also, are you planning on
connecting the load between ground and the 'switch circuit' or between
some power rail and the 'switch circuit'?

Assuming the above doesn't stray too far from your situation, one note
right now is that you should want to keep your PIC drive current as
low as is reasonable, while being able to reliably drive the switch
circuit. I'd target _much_ less than 25mA as reasonable, here.
Something on the order of just a few mA. I don't like the idea of
just seeing the maximum spec of 25mA and merrily planning on using
that as an option.

Jon

 

[default]

Thanks default, I will look at the links you offerd. You mentioned
using MOSFETs, I may not be able to use these because I need to drive
two different base(gate) with less than 25mA from a PIC Output signal
that also goes through a voltage drop (due to a diode) to about 3.8v.
So the MOSFETS, without a transistor amplifier, are out of the
question.

Not trying to push mosfets - but I find they can do some interesting
things at times and save stages of amplification - you have 25 ma of
drive available from the pic, but because the mosfet only needs a
microamp it may work because the voltage out will be higher for lower
current loading.

As some respondants already pointed out - you tend (should) use what
everyone else uses. It is proven, it is in good supply, and is cheap
- and the manufacturers get better at making them - the stated spec
might be 60 volts, for instance, but you may find they all can go to
better than 100 volts and keep working..

Some good ones to use if you can:

2N3906, TO92 PNP
2N3904, TO92 NPN
2N2222 . . . 2N2222A, TO18 metal, PN2222, TO92 NPN
2N2102, TO5 metal NPN
2N2905, TO5 metal PNP
2N3055, ancient and ubiquitous for cheap NPN TO3 metal power device
with a host of work alikes with better specs . . . 2N3771, 2N3772,
2N3773
2N7000 mosfet, TO92
MPSA13, darlington, TO92 NPN
TIPXXX TO220 - and other cases - large selection - both darlingtons
and ordinary bipolar power devices Both Sexes
IRFXXX TO220 ditto

It does look like TO5 and TO3 transistors have/are gone/going out
of favor for cheaper plastic parts Like the TO137, TO147, and TO220
etc.. When in doubt buy the cheapest part that will work or see what
others use.

 

[[email protected]]

Thanks default, I will look at the links you offerd. You mentioned
using MOSFETs, I may not be able to use these because I need to drive
two different base(gate) with less than 25mA from a PIC Output signal
that also goes through a voltage drop (due to a diode) to about 3.8v.
So the MOSFETS, without a transistor amplifier, are out of the
question.

At 25mA, the PIC outputs probably won't be 4.5V (assuming a 5V supply,
about which I'm also not sure.) My guess is that they will be about
25mA*70ohms or 1.75V less than your supply voltage before they meet
your diode. (Actual impedance will vary some and you need to be aware
that you cannot drive all pins at the same 25mA, as there is another
spec that provides an absolute maximum [don't get too close to it]
that limits the total port output.) Or were you thinking that it was
a diode there in the PIC to explain the 3.8V output?

Anyway, I think we understand the following:

Drive signal: PIC output
Number of circuits: 2
Mode of operation: switch (unless I assume far too much here)
Load current: 220mA
Load impendance: unknown, but may have significant L, if relay
Load voltage rail: unknown
PIC Voltage rail: ??, but either 4.5V (calculated from your
assumed diode drop and 3.8V mentioned) or
else 5V (blindly assumed as "normal")

Could you expand the above to provide more precise information? A
highly inductive or highly capacitive load may require other caveats
as compared to a purely resistive load. Also, are you planning on
connecting the load between ground and the 'switch circuit' or between
some power rail and the 'switch circuit'?

Assuming the above doesn't stray too far from your situation, one note
right now is that you should want to keep your PIC drive current as
low as is reasonable, while being able to reliably drive the switch
circuit. I'd target _much_ less than 25mA as reasonable, here.
Something on the order of just a few mA. I don't like the idea of
just seeing the maximum spec of 25mA and merrily planning on using
that as an option.

Jon

Thanks for replying. Using transistors, I was only planning on using at
most 6mA of the 25mA. I dont have the actual spec sheet of the PIC, the
guys I bought it from put a dot over the P/N of it. However they didnt
say 25mA Max, they just said 25mA, so maybe it is a nominal output.
The loads I am driving with the transistors arent large, probobly no
more than a relay load, around 150mA. However I also wanted to drive a
MOSFET from one of the transistors to turn on/off a 556 PWM circuit.
I was thinking using the 2N4401, someone on another group suggested it
to me. The thing is that I am not totally sure because I am a little
weary about the math behind choosing a transistor.

 

[Jim Land]

I may not be able to use these because I need to driveI bet you are not the first person in the world to want to interface a PIC.
Have you looked for circuits that already do what you want to do?

 

[John Popelish]

I was thinking using the 2N4401, someone on another group suggested it
to me. The thing is that I am not totally sure because I am a little
weary about the math behind choosing a transistor.

Lets take a look at the data sheet:
http://www.onsemi.com/pub/Collateral/2N4401-D.PDF

The collector saturation region graph on 6 shows that the collector
saturation voltage falls below .6 volts for a base current of about a
milliamp, and the DC gain curves show the gain falling only a little
between 100 and 200 mA, so you can expect this sort of saturation
voltage at 200 mA with about 2 mA of base current. If you can afford
4 or 5 mA of base current, you should have some margin for a low gain
transistor.